Abstract
Purpose :
It is difficult to predict future myopic shift in pediatric patients. Ocular biometric data is scarce for young children because acquiring keratometry and axial length measurements typically requires general anesthesia. This study leverages data from the Infant Aphakia Treatment Study (IATS) to evaluate the association between axial length elongation during infancy (age 0-1 years) and childhood (age 1-10 years).
Methods :
We used biometric data from the IATS, which collected longitudinal keratometry and axial length measurements on infants (age 0-1) undergoing unilateral cataract surgery. First, we fit a logarithmic regression to model axial length as a function of age. Next, we construct two cohorts: fast infant growers (patients whose axial length change between cataract surgery and 1 year postoperatively exceeded our model's prediction) and slow infant growers (elongation less than model prediction). We then compared the future change in axial length (ages 1 to 5, 5 to 10, and 1 to 10 years) between slow and fast growers using student’s t-test.
Results :
This study included 114 patients (114 phakic eyes, 57 aphakic eyes, 57 pseudophakic eyes). Axial length was fit by a logarithmic function (0.851 * log2(age in days) + 12.851) (Figure 1). The change in axial length between fast growers and slow growers was not significantly different between ages 1-5, 1-10, and 5-10 years (p values 0.53, 0.71, 0.95 respectively) (Figure 2). This finding remained true for sensitivity analyses (1) investigating the non-cataractous fellow eye and (2) comparing the 25th percentile fastest growers with the 25th percentile slowest growers. 50% of infants with fast axial elongation remained fast growers between age 1-10, whereas 60% of infants with slow axial elongation changed to fast growers.
Conclusions :
Axial elongation follows a logarithmic pattern with respect to time. Accelerated axial elongation during infancy did not correlate with accelerated axial elongation during childhood years. Study limitations include a small sample size, inherent noise in EUA measurements, and a biased sample of infants who underwent cataract surgery. The lack of correlation will increase the complexity of building models to predict future refractive error. The underlying biological triggers that accelerate or decelerate axial elongation should be investigated further.
This abstract was presented at the 2023 ARVO Annual Meeting, held in New Orleans, LA, April 23-27, 2023.